Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
Kaiepi how do i make an NQPArray into a perl 6 list of some sort? 03:21
03:25 vrurg left 04:29 skids joined
Kaiepi never mind i was fucking something up 04:35
04:45 skids left 05:30 robertle left 05:33 brrt`` joined 06:03 discord6 left, discord6 joined, discord6 left, discord6 joined 06:06 discord6 left, discord6 joined 06:07 tyil left, Geth left 06:08 discord6 left 06:19 synopsebot_ joined 06:27 Geth joined, tyilanmenyn joined 06:34 tyilanmenyn left 06:36 synopsebot_ left, Geth left, synopsebot left, Geth joined 06:37 travis-ci joined
travis-ci NQP build failed. usev6 'Set language back to perl in travis config 06:37
travis-ci.org/perl6/nqp/builds/532649385 github.com/perl6/nqp/compare/d6d0e...add729b54b
06:37 travis-ci left 06:38 tyilanmenyn joined 06:43 synopsebot joined, discord6 joined, discord6 left, discord6 joined 06:47 Geth_ joined 06:48 tyilanmenyn left 07:00 discord6 left 07:01 Geth left 07:04 Geth joined 07:11 discord6 joined, discord6 left, discord6 joined 07:14 patrickb joined 07:16 Geth left 07:17 discord6 left, Geth joined 07:18 synopsebot_ joined, Geth left, synopsebot_ left, synopsebot left, Geth_ left 07:19 discord6 joined, discord6 left, discord6 joined 07:20 tyilanmenyn joined 07:22 discord6 left 07:23 tyilanmenyn left 07:25 Geth joined, discord6 joined, synopsebot joined, tyilanmenyn joined 08:16 robertle joined 08:33 tyilanmenyn left, synopsebot left 08:34 synopsebot joined 08:35 tyilanmenyn joined 08:39 tyilanmenyn is now known as tyil 08:48 brrt`` is now known as brrt 09:34 brrt left
lizmat .tell jnthn there's still something deeple troublesome about b9f899541b3f153e after a revert I need to do a full reconfigure to get S06-operator-overloading/imported-subs.t not die on missing dependency in Actions.nqp 09:39
yoleaux lizmat: I'll pass your message to jnthn.
jnthn lizmat: Please just revert it 09:41
yoleaux 09:39Z <lizmat> jnthn: there's still something deeple troublesome about b9f899541b3f153e after a revert I need to do a full reconfigure to get S06-operator-overloading/imported-subs.t not die on missing dependency in Actions.nqp
lizmat will do
jnthn But yeah, something is odd if a simple re-build is not enough 09:42
Geth rakudo: 6a60a99fcd | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp
Revert "Fix compile_in_context with an immediate block"

This reverts commit b9f899541b3f153ea73e33269eb53ec34ddf9280.
This was causing strange "Missing dependency" errors in Cro and spectest.
jnthn While I'd suspect b9f899541b3 primarily, it's also a tiny bit suspect that the behavior changes over rebuilds when there was recently a bunch of configure/build changes merged.
Geth roast: 40c3f305d9 | (Elizabeth Mattijsen)++ | S02-types/hash.t
Add tests for R#2865
09:47
synopsebot R#2865 [open]: github.com/rakudo/rakudo/issues/2865 [tests needed] All Junctions behave the same when used to define a hash key
10:09 brrt joined, brrt left
[Tux] Rakudo version 2019.03.1-398-g6a60a99fc - MoarVM version 2019.05-7-gf8743c112
csv-ip5xs0.736 - 0.736
csv-ip5xs-206.063 - 6.326
csv-parser23.184 - 25.142
csv-test-xs-200.431 - 0.472
test7.738 - 8.023
test-t1.718 - 1.734
test-t --race0.792 - 0.807
test-t-2028.669 - 30.835
test-t-20 --race9.306 - 10.082
10:16
nine Woah! nine@sunshine:~/rakudo (master *=)> bash config.status 10:54
config.status: line 1: syntax error near unexpected token `('
config.status: line 1: `/usr/bin/perl Configure.pl "--prefix=/home/nine/rakudo/install" "--backends=moar" "--expand" "moar/perl6-valgrind-m" "--out" "perl6-valgrind-m" "--set-var=MOAR=/home/nine/rakudo/install/bin/moar" "--set-var=ctx_subdir=moar" "--set-var=runner_opts=--execname="@envvar(EXEC)@" --libpath="@envvar(DIR)@" --libpath="@nfp(@envvar(DIR)@/blib)@" --libpath="@nfp(@libdir@/nqp/lib)@"
"@nfp(@envvar(DIR)@/perl6.moarvm)@" --nqp-lib="@nfp(@envvar(DIR)@/blib)@"" $*'
And the reason for the rebuild is not much better: Missing or wrong version of dependency 'gen/moar/Metamodel.nqp' (from '/home/nine/rakudo/lib/CompUnit/Repository/Staging.pm6 (CompUnit::Repository::Staging)') 10:56
Haven't seen those in a long while
Seems like detection of stale precomp files is completely broken? 11:32
patrickb .tell vrurg See here: colabti.org/irclogger/irclogger_lo...-05-15#l99 11:39
yoleaux patrickb: I'll pass your message to vrurg.
nine Boy this makes developing pain
What's the rakudo version I can downgrade to to get it working again? 11:48
patrickb nine: If it's caused by what I suspect then this commit: 292235ce80e8a4d9e54391017fe9780de7d335bd 11:49
nine patrickb: "Update release_guide.pod"? 11:50
patrickb Yes. That's the commit before the build-revamp merge 11:51
nine Ah to downgrade to
patrickb yes :-)
nine Yes, that's much better now. Thanks a lot! 11:55
patrickb yw
jnthn Hmmm, so maybe my fix yesterday wasn't to blame after all 12:19
patrickb .tell vrurg Also notice this: colabti.org/irclogger/irclogger_lo...-05-15#l69 Chances are good it's a common cause for both problems. 12:29
yoleaux patrickb: I'll pass your message to vrurg.
13:22 vrurg joined 13:39 skids joined
Geth Ā¦ problem-solving: AlexDaniel assigned to jnthn Issue What are the intended semantics of <== and <<== github.com/perl6/problem-solving/issues/27 13:49
vrurg jnthn: if you're around, I nudge you as requested. ;) 14:03
yoleaux 11:39Z <patrickb> vrurg: See here: colabti.org/irclogger/irclogger_lo...-05-15#l99
12:29Z <patrickb> vrurg: Also notice this: colabti.org/irclogger/irclogger_lo...-05-15#l69 Chances are good it's a common cause for both problems.
vrurg .tell patrickb got the message. Have seen the problem on my side too. make clean helps, but sure will take care of it. 14:09
yoleaux vrurg: I'll pass your message to patrickb.
patrickb vrurg++ 14:20
yoleaux 14:09Z <vrurg> patrickb: got the message. Have seen the problem on my side too. make clean helps, but sure will take care of it.
patrickb vrurg: So a simple reconfigure fixes those problems? If yes I'm not sure a fix apart from perhaps a notification during make is necessary. 15:00
vrurg: If it's actually that simple, can you ping nine and lizmat to give it a try? 15:02
lizmat I found that nuking all precomp dirs involved, seems to solve the issue 15:11
15:22 patrickb left
vrurg Nuking precomps isn't the way, I think. It seems that changing any of CORE files lead to the problem. Then, yes ā€“ make clean to do correct rebuild, and removal of precomps to allow the code to run. 15:31
.tell jnthn just a reminder about myself. Thanks! 15:44
yoleaux vrurg: I'll pass your message to jnthn.
15:45 vrurg left 15:55 robertle left 16:08 patrickb joined
ugexe I'm not sure nuking precomps is an option 16:15
consider multiple perl6 installs (which would share ~/.perl6), or upgrading an existing perl6. i imagine one of these would not be happy with nuking precomps 16:17
jnthn I'm not sure anyone is proposing that as an actual solution, just observing it helps, which is probably a clue to what's going wrong. 16:18
yoleaux 15:44Z <vrurg> jnthn: just a reminder about myself. Thanks!
ugexe ah, i thought it was being suggested that `make clean` should remove all precomps e.g. ~/.perl6 16:21
16:23 TimToady joined
jnthn .tell vrurg Invitation sent. Welcome! 16:26
yoleaux jnthn: I'll pass your message to vrurg.
lizmat ugexe: my suggestion was to just fix the problems that b9f899541b3f153ea73 caused 16:34
17:15 vrurg joined 17:40 pamplemousse_ joined
vrurg Who do I ask now about getting access to nqp repo and moving a repo under Perl6 organization? 17:46
yoleaux 16:26Z <jnthn> vrurg: Invitation sent. Welcome!
vrurg .tell jnthn Invitation accepted. Thank you!
yoleaux vrurg: I'll pass your message to jnthn.
18:32 dogbert17 joined, pamplemousse_ left
dogbert17 hmm, perl6-valgrind-m is broken all of a sudden 18:33
the message is 'ERROR: Unknown flag --.' 18:35
vrurg dogbert17: the latest master? 18:58
dogbert17 vrurg: yes 19:04
vrurg Is it happening while building/installing or when you run it? 19:06
dogbert17 when I run it 19:07
the last line of perl6-valgrind-m looks like this on my machine, you can see the stray '--' 19:08
valgrind ${MVM_VALGRIND_OPTS} "/home/dogbert/repos/rakudo/install/bin/moar" -- --execname="$EXEC" --libpath="$DIR" --libpath="$DIR/blib" --libpath="/home/dogbert/repos/rakudo/install/share/nqp/lib" "$DIR/perl6.moarvm" --nqp-lib="$DIR/blib" "$@"
vrurg dogbert17: What was your Configure.pl arguments? 19:11
dogbert17 perl Configure.pl --gen-moar --gen-nqp --backends=moar
vrurg Weird. There simply cannot be extra --. 19:12
In your Makefile find $(M_VALGRIND_RUNNER): target, please, and tell what is in the line with --set-var=runner_opts= 19:14
Ignore it, line wrapping in irc client played a trick to me. 19:17
dogbert17 vrurg: found something suspicious?
vrurg It's a smoking gun. Copy/paste error, I think. If you remove that -- it runs as expected, right? 19:19
dogbert17 yep
vrurg Ok, few minutes 19:20
dogbert17 ++vrurg 19:21
Geth rakudo: ee2297b58b | (Vadim Belman)++ | 3rdparty/nqp-configure
Update nqp-configure with jvm fix

jna was missing from java libraries
19:26
rakudo: 9ba51a8b4a | (Vadim Belman)++ | tools/templates/moar/perl6-valgrind-m.in
Remove stray '--' from valgrind runner's moar command line.
vrurg dogbert17: must be ok now.
19:28 robertle joined
dogbert17 it does indeed work, vrurg++ 19:32
19:42 MasterDuke left
Kaiepi i can't figure out what i need to do to get Rakudo::Internals.EVALUATE-FEED to work 19:54
i have the pipeline set up and everything but the stages are still QAST nodes so it doesn't work when i try to run it
do i need to use nqp::getcomp('QAST').as_mast/as_jast/etc. or something? 19:55
lizmat Kaiepi: do you have a gist ? 19:57
Kaiepi lizmat, gist.github.com/Kaiepi/72035a9216a...bd01e47020 19:59
lizmat Kaiepi: I think you're misunderstood 20:00
the nqp bits that used to live in Action, should stay there
Kaiepi oh
lizmat they should just set up an array, which then gets passed to EVALUATE-FEED
Kaiepi an array of what? 20:01
lizmat inside R:I you should only need to know that the @stages is an array of Callables
and the result of the call to R:I.EVALUATE-FEED should be fed to a .STORE method on the pointy end 20:02
hope that makes sense... 20:03
Kaiepi what about cases where .STORE would need to be called in one of the stages like `my @ <== map { $_ ** 2 } <== my @ <== map { .succ } <== 1..10`? 20:04
lizmat if it is a chunk that just receives, you codegen the .STORE on that, and use the result as the source for the next feed sequence 20:07
so in the above example, you would have 2 calls to R:I::EVALUATE-FEED, in my opinion
Kaiepi ohh ok
i think i get what you mean now
vrurg lizmat: Can you guess what's causing 'Missing or wrong version of dependency 'gen/moar/BOOTSTRAP.nqp'? I trying to locate the problem, but narrowing down the search area would be appreciated. 20:08
lizmat .STORE returns the result of the STORE:
m: my %h; dd %h.STORE((a => 42))
camelia Hash %h = {:a(42)}
20:08 ufobat_ joined
lizmat vrurg: I've seen several incarnations of that bug 20:08
after reverting b9f899541b3f153ea73 I did a make clean, and nuked all of the precomp dirs I could find 20:09
ah, and did a reconfigure
vrurg It happens if one of the CORE.setting sources gets changes and then 'make install' fails.
lizmat then it went away fro me
vrurg lizmat: you don't need to reconfigure, just make clean works. 20:10
Unfortuantely, removing precomps works for 'make install' but we can't ask everybody to nuke their precomps on every test run. 20:12
20:12 ufobat__ left
lizmat well.. hopefully not too many people built their rakudo while b9f899541b3f153ea73 was in there 20:12
those are the only one suffering from these problems afaik 20:13
or are you saying a clean checkout now also has that problem ?
vrurg lizmat: Nah, I didn't try the clean one. Ok, lunch time. Got your note on b9f899541b3f153ea73 because I still have it in my history. 20:14
thanks! 20:15
patrickb vrurg: nine might be a good person to ping. He knows a bit about the precomp system. 20:30
Geth roast: b0ab879996 | (Elizabeth Mattijsen)++ | S02-literals/hash-interpolation.t
Add tests for #2825
20:46
lizmat vrurg: welcome !! :-) 20:47
Kaiepi lizmat, now i get "Type check failed in binding to parameter 'item'; expected Any but got QAST::Var (?)" when i run "my @ <== [1,2,3]" 21:23
looks like this now gist.github.com/Kaiepi/72035a9216a...bd01e47020
oops i still need the call to &infix:<,> 21:25
21:29 robertle left 21:32 skids left
Kaiepi oh i need to change Channel.send so it accepts Mu as an argument 21:33
vrurg reversal of b9f899541b3f153ea73 doesn't help. Ok... :( 21:45
timotimo committable6: releases srand(1); my @timings = rand xx 500; @timings .= sort; say @timings[* * 3 div 4, * - * div 100] 23:02
committable6 timotimo, gist.github.com/7520a8753cf6282bbe...404c6ad1d4 23:03
Geth nqp: vrurg++ created pull request #537:
Do the right thing when choosing backends
23:25
vrurg timotimo: are still around? 23:27
23:35 patrickz joined
jnthn vrurg: I'm guessing you already have commit access on the nqp repo? 23:37
vrurg rakudo, but nqp is different organization, isn't it?
jnthn Ah, ok 23:38
Invite sent for that too then
vrurg Thanks!
23:38 patrickb left
vrurg I just have replied to your comment. It seems that you're the only one who can review R#2852 before merging. 23:38
synopsebot R#2852 [open]: github.com/rakudo/rakudo/pull/2852 Add revision 6.e and improve multi-revision support
23:45 patrickz left
jnthn vrurg: Gave what comments I could, but seems they're not really issues :) 23:48
(Based on your replies)
vrurg Ok, I'll prepare roast, do something about hardcoded PREVIEW and merge when ready. Thank you! 23:49
jnthn Nice; thanks!
Geth nqp/master: 5 commits pushed by (Vadim Belman)++ 23:57